A synchronisation of the type mentioned involves each station in the network forming its frequency as a mean value of the frequencies of the incoming links. This is done by measuring the phase differences between incoming link clock signals and the local clock, subsequent to which the phase positions are added or weighted together to provide a control value acting on the clock in the respective exchange via a regulator. The phase detector may be analog or digital, although only a digital phase detector is suitable for use in a time multiplex system.
A phase detector represents a periodic function, i.e. the variation of the faulty signal between zero and the highest deviation corresponding to the phase difference 2.pi. is repeated for each 2.pi. phase difference. This means that large phase jumps occur just when the phase position exceeds or falls below its measuring range. Even if the sum of the phase positions is kept constant in a network, the individual phase positions can move apart and cause phase jumps. This can result in difficulties when the clock frequency rapidly changes, thus possibly causing new phase jumps. Since the entire network is synchronised, such a phase jump can cause the whole network to hunt.